View from Mt. Washington, NH
A photo I took from Mt. Washington, NH, nine miles east of Bretton Woods.

Isaac Slavens

Throughout the 21st century, economic stability has become dependent on ever-growing sovereign debt.

This dependence, notably in the United States (charted here), accelerated in 2008 and again in 2020. These featured charts summarize why I am focused on this issue, given the clear divergence.

This site examines the nature and consequences of this economic environment in further detail, with a focus on sovereign debt. It is my primary interest right now.

Read the articles below ↓

Switch between the two charts:

Response of the US primary balance to lagged debt by era, with 95% confidence intervals: 0.029 for FY1950-1974, 0.059 for FY1975-2000, and -0.038 for FY2001-2025 Rolling 30-year estimate of the US primary balance's response to lagged debt with 95% confidence band, above the annual output gap; FY2025 estimate -0.067, p = 0.004

Data: FRED. Made in Python with Claude. The response of the US primary balance to lagged debt (g), estimated separately for FY1950–1974, 1975–2000 and 2001–2025 in a single regression, controlling for the output gap and COVID. Following Bohn (1998), g > 0 means primary surpluses rise with debt, a sufficient condition for sustainability. Bars are 95% confidence intervals.

Full specification

A single OLS regression over FY1950–2025: PBt = αk + β·Gapt + Σ gk·(Dt−1 × Ek,t) + δ·Covidt + εt, where Ek equals 1 in era k (FY1950–1974, 1975–2000, 2001–2025) and 0 otherwise.

PB is the primary balance and D is debt held by the public, both in % of GDP. Gap is the output gap against CBO potential GDP, and Covid is a FY2020–21 dummy. Each era has its own intercept αk, so gk measures the debt response within that era, while β and δ are shared across eras. Intervals use Newey-West (HAC, 2 lags) standard errors.

Data: FRED. Made in Python with Claude. The same debt response (g), re-estimated over each 30-year window and plotted at the window's final fiscal year, controlling for the output gap and COVID. Below zero, primary surpluses no longer rise with debt. Shading is the 95% confidence interval; the lower panel shows the annual output gap. Latest window, FY1996–2025: g = −0.067 (p = 0.004).

Full specification

Each point is an OLS estimate of PBt = α + β·Gapt + g·Dt−1 + δ·Covidt + εt over the 30 fiscal years ending in that year, using data from FY1950–2025.

PB is the primary balance (federal surplus plus interest outlays, % of GDP), D is debt held by the public (% of GDP), Gap = 100 × (real GDP − CBO potential GDP) / potential GDP, and Covid is a FY2020–21 dummy. Following Bohn (1998), g > 0 means primary surpluses rise with debt, a sufficient condition for sustainability. Intervals use Newey-West (HAC, 2 lags) standard errors.

"Unless we think of our civilization as something that's going to stretch out into the future, we won't do the job of living well today."

Hyman Minsky


On AI Usage and Accountability

I use AI to code this site and to assist in research, but not to write anything for me. All writing and reasoning is independent. I welcome any attempt to verify or disprove this commitment.

Isaac Slavens

Nothing on this website is financial advice.